LinuxServer Changes:
Rebase to alpine 3.15.
mylar3 Changes:
Improvements
IMP: Refactor version checking entirely:
- Minimum time allowed between startups when check_on_startup is enabled
- better GUI notification - always present when commits are available and app is out-dated
- docker notification - display notification of new Mylar releases (not new docker releases)
- proper branch display - regardless of ini setting, proper display in GUI (no more None values)
- branch name will now be displayed again regardless of source/git/docker installation (master branch releases only)
IMP: Add unwatched series from Arc to watchlist option on arc detail page
IMP: config.ini addition: pack_priority for DDL downloads on a global scale (True = packs prioritise over individual issues, False issues over packs (the default))
IMP: OPDS endpoint configuration available in OPDS GUI section of configuration for custom opds endpoints
IMP: Add api hit / provider calculation to database entries (getting ready for future support)
IMP: Series folder collision will now be run on each series directory creation to try to minimise folder hammering (will either append $Type or $VolumeY/N depending on which is available for specific series)
IMP: When checking pull for new issues will now use the last time series was updated in calculation of marking issue as Wanted instead of current date (if Mylar was off for a period of time and passed the current pull, or previous pulls would not mark)
IMP: Add GNUv3 blurb to search_filer module
IMP: When DB updater is running, run it thru the Refresh Queue instead of an independent thread, so that multiple refeshes don't run simultaneously causing possible collisions
IMP: Add Incomplete section to Wanted tab / Upcoming section. Issues displayed within do not have CV information present as of yet (usually for new issues)
IMP: Added regenerateCovers API endpoint to allow for regeneration of series covers in cache folder see here
IMP: Added reason and additional information to Mismatch table on Wanted page / Upcoming section to try to briefly explain why the mismatch
IMP: Minimum_py_version added as a global variable to indicate minimum python version required (set to 3.8.1 currently)
IMP: Warning will only display if version running Mylar is less than minimum required
IMP: Colour coded warning on config page to emphasis importance (hopefully) & python version line colour-coded red
Fixes
FIX: Experimental search would throw traceback on searching for print versions of issues with no file name
FIX: If file already exists when DDL goes to download and resume isn't enabled, would cause traceback
FIX: Variables not formatted in a logger line within the failed module
FIX: If file already exists during post-processing without the [ISSUEID] (but has it within the original filename) in the same location (ie. DDL) - would throw a FileExistsError and halt post-processing queue
FIX: Incorrect variable reference during maintenance cleanup
FIX: Depreciated threading.currentThread & threading.isAlive calls
FIX: Schedulers will display proper status in GUI on startup (based on prior load state to allow for persistent status')
FIX: Suppress unnecessary SSE event warning, and some additional logging
FIX: Alpha-index would throw a 500 error on a null ComicName if add/refreshing said Comic failed at a bad state
FIX: Remove unused module scheduler.py since we've been using apscheduler for years
FIX: logging error on non-EN systems during maintenance resultling in non-operational bootup
FIX: Remove Read Sync and Clear File Cache from StoryArc detail pages as they're non-working
FIX: indicate Data directory check on startup, instead of saying comic directory (misleading)
FIX: Schedulers would run even though config settings would have them disabled
FIX: Loading previous states from db would not take into account status' prior to the update
FIX: monitor_scheduler was mislabelled as folder_scheduler which probably caused some problems
FIX: When folder monitor was run (whether enabled or not) and no folder was set, would traceback error and lock up the thread
FIX: If job is in a Paused status, there is no Next Runtime displayed in the GUI
FIX: Force job will now only run the job once instead of inadvertently making it a scheduled job (even if it was Paused prior to the Force job)
FIX: (#1057) StoryArc Refresh option would not work when apostrophes in Story ArcName
FIX: (#1055) (Logs page) page numbering not rendering correctly on large datasets
FIX: Update logo for Aardvark
FIX: Annuals containing [issueid] in the filename would fail to post-process
FIX: Post-processing completion for packs include series attached to if possible
FIX: When nzb downloads failed to download, failed retry would not cycle properly and re initiate causing hanging queue
FIX: GUI will now display a notification when issues fail to download when failed handling is enabled
FIX: Try to improve basic pack matching with DDL
FIX: Correct some incorrect checks for True/False values in weeklypull
FIX: When scanning files on page load, subdirectory paths would not be fully recognized if more than one level deep
FIX: Remove some more spam / unnecessary logging
FIX: incorrect resolving of %s when doing queries against GC (would do a query for * instead of specific query)
FIX: SSE display during DBUpdate would have a None for year
FIX: Remove year_check logging during file scanning
FIX: Storyarc detail options (grab and retry) would not work with story arcs with an ' in the title
FIX: (Post-Processing) Fixed DDL (or files tagged with the [issueid]) path location problems
FIX: (Post-Processing) One-offs / storyarcs in conjunction with using [issueid] file tag
FIX: (Post-Processing) Tidyup after Post-Processing would fail to cleanup remnants with a path error
FIX: Scheduler displaying as Running when it's not
FIX: default_dates ini setting would not be honoured when trying to set default view on series detail page for Store / Cover dates
FIX: Check version results could not be seen in default interface due to white color
FIX: Fix upcoming table in Wanted page css layout
FIX: Apparently Snatche isn't a proper word
FIX: Torrent torznab would not be identified as such causing problems with searching, downloading, etc.
FIX: Try to properly generate an id from torrent url (alternate method)
FIX: Remove 32P from config page
FIX: Mismatched status on pull-list clickable back to Upcoming / Mismatched table to show possible explanation
FIX: DB Updater adding identical series during check
FIX: DB Updater would not fire off after queuing up items
FIX: Typo when parsing [issueid]
FIX: Additional logging during path conversions with ddl
FIX: Watch on weeklypull would not add if publisher contained an ampersand
FIX: (#1085) packs in RSS search results would cause key error
FIX: CV titles that start with a space(?) would cause filenaming inconsistencies (as well as display order problems)